Intelligent Analysis System and Method for Mammogram Images Based on Multi-Scale Deep Learning

Abstract

The present invention relates to the technical field of intelligent medical image analysis systems, and specifically relates to an intelligent mammogram image analysis system and method based on multi-scale deep learning, including: an image acquisition module for acquiring mammogram images; a calcification detection module connected to the image acquisition module for intelligently detecting calcification points in mammogram images based on a multi-dimensional feature matrix and an adaptive feature filter, and generating confidence scores and dynamic segmentation thresholds for the calcification points; a calcification segmentation module connected to the calcification detection module for segmenting the calcification point regions based on the dynamic segmentation thresholds of the calcification detection module; a calcification quantification module connected to the calcification segmentation module for extracting quantitative features of the calcification points based on the segmentation results of the calcification segmentation module. By introducing feature extraction and feature fusion at different scales, the sensitivity to microcalcification points is significantly improved, and at the same time, the problem that traditional single-scale models cannot accurately identify calcification features at different scales is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent medical image analysis systems, and particularly to an intelligent analysis system and method for mammography images based on multi-scale deep learning. Background Art

[0002] Currently, the screening and early diagnosis of breast cancer increasingly rely on the intelligent analysis of mammography images. Calcification points, as early lesion features of breast cancer, have high diagnostic value. However, due to the diversity of the size, density, and shape of calcification points in mammography images, traditional detection methods based on single-scale feature extraction are prone to interference from background noise, morphological differences, and densely distributed areas in complex breast tissues. These methods mostly use single-scale convolutional neural networks (CNNs) or fixed-threshold segmentation, and cannot accurately identify complex calcification features, resulting in technical problems such as high false-negative rates, high false-positive rates, and low segmentation accuracy.

[0003] In existing single-scale convolutional network models, the detection of calcification points mainly relies on feature extraction of a single scale or direction. This method is often ineffective for tiny or densely distributed calcification points. At the same time, traditional methods lack a flexible feature weighting and segmentation threshold adjustment mechanism, and are prone to misidentifying non-calcified regions in mammography images as calcification points, thus increasing the false-positive rate. Fixed segmentation thresholds usually cannot adapt to calcification points of different densities and sizes. Especially when segmenting calcification points in dense areas, due to the blurred boundaries between adjacent calcification points, the segmentation results are inaccurate and difficult to meet clinical needs.

[0004] In addition, traditional methods lack adaptability in feature extraction and weighting, and fail to assign dynamic weights to calcification points according to the importance of features of different scales and directions. Existing feature extraction methods usually do not have sufficient ability to suppress background noise, resulting in a relatively high false-positive rate when detecting tiny calcification points. Due to the lack of an adaptive feature filter, the system performs poorly in processing complex backgrounds and high-noise environments, limiting the sensitivity and accuracy of calcification point detection. In addition, the computational efficiency of existing methods is relatively low, and it is difficult to quickly process a large number of mammography images, which is not conducive to practical clinical applications. Summary of the Invention

[0005] Aiming at the problems of the above-mentioned existing technologies, the present invention provides an intelligent analysis system and method for mammography images based on multi-scale deep learning, aiming to achieve high-precision calcification point detection through the coordinated cooperation of multiple modules, combined with multi-scale feature extraction and dynamic segmentation threshold adjustment.

[0006] The present invention proposes an intelligent analysis system for mammography images based on multi-scale deep learning, including:

[0007] An image acquisition module for acquiring mammogram images;

[0008] A calcification detection module, connected to the image acquisition module, for intelligently detecting calcification points in mammogram images based on a multi-dimensional feature matrix and an adaptive feature filter, and generating a confidence score and a dynamic segmentation threshold for the calcification points;

[0009] A calcification segmentation module, connected to the calcification detection module, for segmenting the calcification point region based on the dynamic segmentation threshold of the calcification detection module;

[0010] A calcification quantification module, connected to the calcification segmentation module, for extracting quantitative features of the calcification points based on the segmentation result of the calcification segmentation module.

[0011] Preferably, the image acquisition module performs image preprocessing on the mammogram image, and the image preprocessing includes denoising and image enhancement to improve the clarity of the image and the detection accuracy of the calcification points.

[0012] Preferably, the calcification detection module includes:

[0013] Steps for constructing a multi-dimensional feature matrix for extracting calcification point features in mammogram images from different directions and scales to generate a feature matrix M d,s (x, y), where d is the direction dimension, s is the scale dimension, and (x, y) is the pixel point coordinate, and the element M d,s (x, y) in the feature matrix represents the feature value of the pixel point (x, y) at direction d and scale s;

[0014] Using an adaptive feature filter to perform feature screening on the feature matrix M d,s (x, y) to obtain a denoised feature matrix F d,s (x, y), and the adaptive feature filter removes low-intensity background noise based on a set feature threshold T;

[0015] Calculating an adaptive feature weight W d,s , for assigning weights to calcification point features in multi-scale features, where α is a direction adjustment parameter and β is a scale adjustment parameter;

[0016] Generating a calcification point confidence score C(x, y) based on the adaptive feature weight for quantifying the accuracy of the detection result;

[0017] Generating a dynamic segmentation threshold λ(x, y) based on the confidence score for the segmentation operation of the calcification points.

[0018] Preferably, in the steps of constructing the multi-dimensional feature matrix, the generation method of the feature matrix is:

[0019]

[0020] where f(x, y) is the pixel intensity value in the mammogram, and g d,s (x, y) is the two-dimensional feature basis function in the direction d and scale s, which is used to extract the corresponding calcification point feature information.

[0021] Preferably, the adaptive feature filter filters the feature matrix M d,s (x, y) to obtain the denoised feature matrix F d,s (x, y), where

[0022] F d,s (x, y) = M d,s (x, y) · δ(x, y),

[0023] and δ(x, y) is the filtering function, δ(x, y) = 1 when M d,s (x, y) > T, otherwise δ(x, y) = 0. The T represents the feature threshold, which is used to filter the low-intensity background noise.

[0024] Preferably, the adaptive feature weight W d,s is calculated based on the denoised feature matrix F d,s (x, y) to determine the weight distribution of the calcification point in the multi-scale features, where

[0025]

[0026] The α represents the direction sensitivity adjustment parameter, and the β represents the scale sensitivity adjustment parameter, which are used to adjust the influence of different directions and scales on the calcification point.

[0027] Preferably, the calcification point confidence score C(x, y) is generated based on the adaptive feature weight W d,s and the calcification point detection result S(x, y), which is used to quantify the accuracy of the calcification point detection result, where

[0028]

[0029] The γ is the confidence smoothing coefficient, S(x, y) = 1 indicates that a calcification point is detected, and S(x, y) = 0 indicates that no calcification point is detected.

[0030] Preferably, the dynamic segmentation threshold λ(x, y) is dynamically adjusted based on the calcification point confidence score C(x, y), which is used for the precise segmentation of the calcification point region, where

[0031] λ(x, y) = θ · C avg · (1 + C(x, y)),

[0032] The C avg is the average confidence of the breast mammography image, and θ is a threshold adjustment factor, which is used to adjust the segmentation threshold to adapt to calcification points of different densities and shapes.

[0033] The intelligent analysis method of mammary gland mammography images based on multi-scale deep learning based on the system comprises the following steps:

[0034] Generate a multidimensional feature matrix M based on the mammography image d,s (x, y); Use an adaptive feature filter to filter the feature matrix to obtain a denoised feature matrix, and calculate the adaptive feature weight W of the feature matrix d,s ; Generate a calcification point confidence score C(x,y) based on the feature weight; Generate a dynamic segmentation threshold λ(x,y) based on the confidence score; Segment the calcification point area based on the dynamic segmentation threshold; Extract quantitative features of the calcification point.

[0035] Preferably, the mammographic target image is subjected to image preprocessing after being obtained, and the image preprocessing includes denoising and enhancement to improve the quality of image input and enhance the detection of calcification point features.

[0036] The beneficial technical effects brought by the technical solution of the present invention are:

[0037] The system of the present invention includes an image acquisition module, a calcification detection module, a calcification segmentation module and a calcification quantification module. The modules are connected consistently through data flow and information flow to form a highly coordinated closed-loop system structure, thereby ensuring the overall stability and detection accuracy of the system.

[0038] The multi-scale feature pyramid method proposed in the present invention significantly improves the sensitivity to tiny calcification points in the calcification detection module by introducing feature extraction and feature fusion at different scales, while overcoming the problem that the traditional single-scale model cannot accurately identify calcification features at different scales. On the basis of calcification point feature extraction, the present invention adopts an adaptive feature filter to remove background noise and retain high-intensity calcification point information, further suppressing false positives. In particular, in terms of feature weighting, the system generates adaptive feature weights based on the importance of features in different directions and scales, ensuring that the system can accurately weight the features of calcification points at multiple scales and angles, greatly enhancing the robustness of calcification point detection.

[0039] To improve the accuracy of segmentation, the calcification detection module of the present invention dynamically generates a segmentation threshold based on the confidence score of calcification points to adapt to the characteristics of calcification points with different densities and morphologies. The dynamic segmentation threshold solves the problem of easy boundary blurring in the detection of dense calcification regions by a fixed threshold, ensuring more accurate boundary localization of the calcification region by the system. At the same time, the present invention designs an effective data processing flow in the coordination of multiple modules and the closed-loop information flow. A strict logical closure is formed between the modules in terms of information transmission and input-output alignment, enabling the entire system to ensure the detection accuracy while efficiently processing image data.

[0040] Macroscopically, under the synergistic effect of multiple modules of the system of the present invention, high-precision detection, segmentation, and quantitative analysis of calcification points are achieved by using multi-scale feature extraction, adaptive feature weighting, and dynamic threshold segmentation. The tight coupling of multiple modules and the smooth connection of data streams significantly improve the detection performance of the system and achieve higher accuracy in complex backgrounds. Brief Description of the Drawings

[0041] Figure 1 It is the overall logical block diagram of the system of the present invention.

[0042] Figure 2 It is the core algorithm logical block diagram of the present invention.

[0043] Figure 3 It is the segmentation effect diagram of microcalcifications and breast masses in the left breast mammogram of a patient with invasive ductal carcinoma. (a) The craniocaudal (CC) view shows microcalcifications with focused foci (indicated by thin arrows) and an irregular round mass (indicated by thick arrows). (b) The suspicious mass is automatically outlined within the red curve. (c) The segmented microcalcifications detected in (b) are used to characterize the features. Detailed Description of the Preferred Embodiments

[0044] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following combines the drawings and preferred embodiments to detail the specific implementation manner, structure, features, and their effects as follows. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0045] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0046] Refer to Figures 1-3, the present invention provides an intelligent analysis system and method for mammography images based on multi-scale deep learning. The system of the present invention includes an image acquisition module 1, a calcification detection module 2, a calcification segmentation module 3, and a calcification quantification module 4. The image acquisition module 1 is used to acquire mammography images and transmit them to the calcification detection module 2. The calcification detection module 2 detects calcification points in the mammography images based on a multi-dimensional feature matrix and an adaptive feature filter, and generates a confidence score for the calcification points and a dynamic segmentation threshold. After receiving the segmentation threshold output by the calcification detection module 2, the calcification segmentation module 3 achieves precise segmentation of the calcification points by adjusting the segmentation accuracy of the calcification point regions. Finally, the calcification quantification module 4 extracts quantitative features of the calcification points, such as area and quantity, based on the segmented calcification regions, providing a strong basis for subsequent medical diagnosis.

[0047] Preferably, the image acquisition module 1 can perform appropriate preprocessing operations based on the specific quality of the mammography images. For example, the image can be enhanced by an adaptive histogram equalization method to improve the contrast, making the calcification points more visible in the image, thus being more conducive to the subsequent detection by the calcification detection module 2. Such a modular structure ensures the consistency of input and output between modules, enabling efficient and accurate transmission of image information and forming an integrated intelligent analysis system.

[0048] Preferably, the image acquisition module 1 can also include functions for denoising and enhancing the mammography images. The denoising process can be based on Gaussian filtering or adaptive filtering methods to smooth out the high-frequency noise in the image, thereby improving the clarity of the image. For the enhancement process, an adaptive contrast enhancement algorithm can be used to enhance the saliency of the calcification points, making the calcification regions clearer in the image, so as to be able to detect and analyze the calcification point features more accurately. Such preprocessing methods not only effectively improve the input quality of the image but also ensure the detection performance of the system in a complex background environment.

[0049] In an embodiment of the present invention, for the calcification detection module 2 of the present invention, first, the step of constructing a multi-dimensional feature matrix is used to extract the calcification point features in the mammography images from different directions and scales, generating a feature matrix M d,s (x,y), where d is the direction dimension, s is the scale dimension, and (x,y) is the pixel point coordinate. The element M d,s (x,y) in the feature matrix represents the feature value of the pixel point (x,y) in the direction d and scale s;

[0050] Use an adaptive feature filter to perform feature screening on the feature matrix M d,s (x,y) to obtain a denoised feature matrix F d,s(x, y), and the adaptive feature filter removes low-intensity background noise based on a set feature threshold T;

[0051] Calculate the adaptive feature weight W d,s , which is used to allocate weights to calcification point features in multi-scale features, where α is the direction adjustment parameter and β is the scale adjustment parameter;

[0052] Generate a calcification point confidence score C(x, y) based on the adaptive feature weight, which is used to quantify the accuracy of the detection result;

[0053] Generate a dynamic segmentation threshold λ(x, y) based on the confidence score for the segmentation operation of calcification points. The multi-scale feature matrix designed in this way effectively covers calcification features of different sizes and directions, enabling the system to accurately identify calcification points in various situations, especially showing excellent detection effects for microcalcification features.

[0054] Preferably, in the step of constructing the multi-dimensional feature matrix, the generation method of the feature matrix is:

[0055]

[0056] where f(x, y) is the pixel intensity value in the mammogram, and g d,s (x, y) is the two-dimensional feature basis function in the direction d and scale s, which is used to extract the corresponding calcification point feature information.

[0057] Next, the adaptive feature filter mainly removes background noise according to the feature value threshold T, leaving a high-intensity calcification feature area.

[0058] Preferably, in the process of generating the feature matrix of the present invention, the feature value threshold T will be automatically adjusted according to the changes in image brightness and contrast. Generally, the threshold T can be taken as 1.5 to 2.0 times the average gray level of the image to ensure effective removal of low-intensity noise. The filtered feature matrix F d,s is then expressed as

[0059] F d,s (x, y) = M d,s (x, y)·δ(x, y),

[0060] and δ(x, y) is the filtering function. When M d,s (x, y) > T, δ(x, y) = 1; otherwise, δ(x, y) = 0. The T represents the feature threshold, which is used to filter low-intensity background noise. This design can significantly reduce the false detection situation in calcification point detection and improve the signal-to-noise ratio of calcification features.

[0061] In the embodiment of the present invention, the adaptive feature weight W d,sThe calculation method is based on the filtered feature matrix F d,s , and is used to allocate weights between features of different scales and directions. The specific formula is:

[0062]

[0063] The α represents the direction sensitivity adjustment parameter, and the β represents the scale sensitivity adjustment parameter, which are used to adjust the influence of different directions and scales on calcification points. They are usually set to 1.0 and 1.5 respectively to achieve a reasonable weight distribution among features. Through the adaptive calculation of feature weights, the system can automatically focus on the most representative calcification features and further improve the recognition effect of calcification points while suppressing background interference.

[0064] In the embodiment of the present invention, based on the adaptive feature weight W d,s , the calcification detection module 2 generates a confidence score C(x, y) for the calcification points, and its formula is as follows:

[0065]

[0066] The γ is the confidence smoothing coefficient, S(x, y)=1 indicates that a calcification point is detected, and S(x, y)=0 indicates that no calcification point is detected. Among them, γ is usually set between 0.8 and 1.0 to smooth the scoring result. The confidence score effectively quantifies the detection accuracy of calcification points by combining feature weights and detection results.

[0067] In the embodiment of the present invention, the system generates a dynamic segmentation threshold λ(x, y) based on the confidence score C(x, y) to accurately segment the calcification point region. Specifically, the segmentation threshold formula is:

[0068] λ(x, y)=θ·C avg ·(1 + C(x, y))

[0069] where θ is the threshold adjustment factor, usually set in the range of 0.5 to 2.0, and C avg represents the average confidence value of the image. This dynamic threshold can automatically adapt to the morphological and density changes of calcification points, making the segmentation result more accurate, especially suitable for dense calcification regions.

[0070] In the embodiment of the present invention, after receiving the segmentation threshold from the calcification detection module 2, the calcification segmentation module 3 segments the calcification point region according to the dynamic segmentation threshold λ(x, y). The segmentation process can adjust the segmentation accuracy of each region according to the confidence of the calcification points to ensure the accurate coverage of the boundary of the calcification region, thus avoiding false detection and missed detection.

[0071] In an embodiment of the present invention, an intelligent analysis method for mammogram images based on multi-scale deep learning is also disclosed, including the following steps:

[0072] Generate a multi-dimensional feature matrix M based on the mammogram image d,s (x, y); filter the feature matrix using an adaptive feature filter to obtain a denoised feature matrix, and calculate the adaptive feature weight W of the feature matrix d,s ; generate a calcification point confidence score C(x, y) based on the feature weight; generate a dynamic segmentation threshold λ(x, y) based on the confidence score; segment the calcification point region based on the dynamic segmentation threshold; extract the quantitative features of the calcification points. The step-by-step implementation of the above method enables the system to further extract the quantitative features of the calcification points after accurately segmenting the calcification point region, thereby realizing intelligent mammogram image analysis.

[0073] Preferably, the mammogram image is first preprocessed by denoising and contrast enhancement after acquisition. The contrast enhancement operation can preferably use an adaptive contrast enhancement technique to significantly highlight the calcification features, thereby ensuring the detection accuracy of the system in a complex background environment.

[0074] To verify the superiority of the intelligent analysis system and method for mammogram images based on multi-scale deep learning of the present invention, publicly available professional datasets such as the Digital Data base for Screening Mam mography (DDSM) are used for testing. The DDSM dataset contains a large number of mammogram images with detailed annotations and is a standard test set for breast calcification point detection and analysis, which is widely used in the research of computer-aided detection and diagnosis of breast lesions. In this comparative test, 200 mammogram images of different types and characteristics are selected, and tests are carried out on key indicators such as the accuracy of calcification point detection, the precision of segmentation, the accuracy of quantitative feature extraction, and the calculation efficiency of the system, so as to verify the practical application effect of the method of the present invention in the analysis of complex mammogram images.

[0075] In an embodiment of the present invention, a calcification point feature extraction based on a multi-scale feature pyramid, an adaptive adjustment of the filter, and a dynamic segmentation threshold mechanism are used to process the DDSM dataset. For each mammogram image, through the adjustment and optimization of the following key parameters, accurate identification and quantitative analysis of the calcification points are achieved:

[0076] 1. Scale range of the feature pyramid: Perform multi-layer pooling at scales of 1x1, 2x2, 3x3, 6x6, and 7x7.

[0077] 2. Adaptive feature filter threshold T: Set to 1.5 times the average gray value of the image to filter out background noise.

[0078] 3. Feature weight parameter W d,s : It is adaptively adjusted according to the different scales of calcification points to balance the contributions of features at each scale to the detection of calcification points.

[0079] 4. Confidence score smoothing coefficient γ: Set to 0.9 to smooth the confidence scores of calcification point detection.

[0080] 5. Dynamic segmentation threshold adjustment coefficient θ: Set to 1.5 to ensure accurate segmentation of the calcification point region.

[0081] In this embodiment, the above parameters and processes are applied to the DDSM dataset, and high-precision calcification point detection results are obtained.

[0082] In the comparative example, a traditional calcification point detection method based on a single-scale convolutional neural network (CNN) is used. When tested in the DDSM dataset, the traditional CNN method is prone to missing small calcification points in the image, and it fails to adaptively adjust in terms of feature weight assignment, noise filtering, etc. In addition, the comparative example does not use a dynamic segmentation threshold, but uses a fixed threshold, which leads to a high false positive rate and inaccurate segmentation in the dense calcification region.

[0083] The test criteria and methods are as follows:

[0084] The following key indicators are mainly examined in the test:

[0085] Detection accuracy (Accuracy): The ratio of the number of correctly detected calcification points to the total number of calcification points.

[0086] Segmentation precision: The overlapping ratio of the segmented calcification point region to the actual region.

[0087] False positive rate (False Positive Rate, FPR): The false detection ratio of detecting non-calcification points.

[0088] Computing efficiency (Processing Time per Image, PTI): The processing time per image, in seconds (s).

[0089] Among them, the detection accuracy and segmentation precision are calculated by the standard overlap rate of calcification point detection: Considering the overlap degree between the detected calcification point region and the labeled region exceeding 70% as a successful detection, and the false positive rate is obtained through the statistical analysis of the false detection rate. Each detection index is carried out under the same hardware conditions to ensure the fairness of the comparison results.

[0090] The detection results are shown in the following table:

[0091] Index Embodiment of the present invention Comparative example Detection accuracy rate (%) 94.8 82.3 Segmentation accuracy rate (%) 91.5 75.4 False positive rate (%) 4.3 12.8 Computing efficiency (PTI, s) 0.9 1.7

[0092] As can be seen from the detection results, the present invention is significantly superior to the comparative example in terms of detection accuracy, segmentation accuracy, and false positive rate. First of all, in terms of detection accuracy, through the application of the multi-scale feature pyramid, the present invention effectively covers the calcification point features of different scales and directions, making the detection accuracy of calcification points reach 94.8%, showing a significant improvement compared with 82.3% of the comparative example. Especially in terms of segmentation accuracy, the dynamic segmentation threshold adjustment mechanism of the present invention effectively improves the accuracy of calcification region segmentation, making the segmentation accuracy reach 91.5%, which is 16.1 percentage points higher than that of the comparative example, proving the superiority of adaptive dynamic segmentation. In addition, the significant decrease in the false positive rate (only 4.3%) indicates that the adaptive feature filter of the present invention plays a significant role in reducing the influence of background noise. Finally, due to the optimization of the multi-scale structure and weight adjustment, the calculation efficiency of the present invention is also superior to the traditional method, and the processing time per image is shortened to 0.9 seconds, which is faster than the comparative example method. This embodiment verifies the superiority of the present invention in mammogram image analysis.

[0093] Although the specific implementation manners of the present invention are described above in conjunction with the accompanying drawings, it is not a limitation on the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solutions of the present invention, various modifications or deformations that can be made by those skilled in the art without creative efforts are still within the protection scope of the present invention.

Claims

1. Intelligent analysis system of mammary gland mammography images based on multi-scale deep learning, characterized by: include: An image acquisition module, used for acquiring breast mammography images; A calcification detection module, connected to the image acquisition module, is used to intelligently detect calcification points in the mammary gland mammography image based on a multidimensional feature matrix and an adaptive feature filter, and to generate a confidence score and a dynamic segmentation threshold for the calcification points; A calcification segmentation module, connected to the calcification detection module, for segmenting the calcification point area based on the dynamic segmentation threshold of the calcification detection module; A calcification quantification module, connected to the calcification segmentation module, for extracting quantitative features of calcification points based on the segmentation results of the calcification segmentation module; The calcification detection module comprises: Steps for constructing a multidimensional feature matrix to extract features of calcification points in mammography images from different directions and scales and generate a feature matrix ,in is the direction dimension, is the scale dimension, is the pixel coordinate, the element in the feature matrix Indicates in the direction and scale The pixel below The characteristic value of Use adaptive feature filter to filter the feature matrix Perform feature screening to obtain the denoised feature matrix The adaptive feature filter is based on the set feature threshold Remove low-intensity background noise; Calculating adaptive feature weights , used to assign weights of calcification features in multi-scale features; Generating a calcification point confidence score based on the adaptive feature weight , used to quantify the accuracy of detection results; Generating a dynamic segmentation threshold based on the confidence score , used for the segmentation of calcification points; In the step of constructing a multidimensional feature matrix, the feature matrix is ​​generated in the following manner: ， in is the pixel intensity value in the mammography image, For the direction and scale The two-dimensional characteristic basis function on is used to extract the corresponding calcification point feature information; The adaptive feature filter is applied to the feature matrix Filter to obtain the denoised feature matrix ,in ， and is the filter function, when hour ,otherwise , represents the feature threshold, which is used to filter low-intensity background noise; The adaptive feature weight Based on the denoised feature matrix Calculation is performed to determine the weight distribution of calcification points in multi-scale features, where ， Represents the directional sensitivity adjustment parameter, Represents the scale sensitivity adjustment parameter, which is used to adjust the impact of different directions and scales on calcification points; The calcification point confidence score Based on the adaptive feature weight and calcification detection results Generate, used to quantify the accuracy of calcification point detection results, where ， Said is the confidence smoothing coefficient, Indicates that calcification points have been detected. Indicates that no calcification was detected; The dynamic segmentation threshold Calcification-based confidence score Dynamic adjustment is performed to accurately segment the calcification area. ， Said is the average confidence of the mammographic image, is the threshold adjustment factor, which is used to adjust the segmentation threshold to adapt to calcification points of different densities and shapes.

2. The intelligent analysis system for mammographic target images based on multi-scale deep learning according to claim 1, characterized in that: The image acquisition module performs image preprocessing on the breast molybdenum target image, and the image preprocessing includes denoising and image enhancement to improve the clarity of the image and the detection accuracy of the calcification point.

3. A method for intelligent analysis of mammographic images based on multi-scale deep learning based on the system according to any one of claims 1 to 2, characterized in that: The following steps are involved: Generate a multidimensional feature matrix based on the mammography image The feature matrix is ​​filtered using an adaptive feature filter to obtain a denoised feature matrix, and an adaptive feature weight of the feature matrix is ​​calculated. Generate a calcification point confidence score based on the feature weights Generating a dynamic segmentation threshold based on the confidence score The calcification point area is segmented based on the dynamic segmentation threshold; and the quantitative features of the calcification point are extracted.

4. The method for intelligent analysis of mammary gland molybdenum target images based on multi-scale deep learning according to claim 3, characterized in that: The mammographic target image is preprocessed after being obtained, and the image preprocessing includes denoising and enhancement to improve the quality of image input and enhance the detection of calcification point features.

Images